Label-free polar metabolite quantification for untargeted metabolomics

总结 NIH化合物鉴定开发核心（CIDC）的主要重点是使用非靶向的 代谢组学不仅可以识别新的代谢物，还可以促进和改善已知代谢物的识别。 代谢物。此外，CIDC的任务是促进准确性，再现性和实验室间 代谢组学数据的比较。促进可重复性、提高可比性和增强 代谢物鉴定的可信度是提高代谢物的定量，特别是对于非靶向代谢物， 代谢组学事实上，正如非靶向NMR研究经常显示的那样， 特定代谢物的浓度可以“纳入”或“排除”试验性鉴定。例如，如果代谢物信号 暂时被鉴定为犬尿烯酸，但测得的浓度是100倍以上 那么，该临时标识一定是不正确，因此，“被排除”。传统复合 代谢组学中的定量（尤其是绝对定量）仅限于靶向代谢组学 而非靶向方法主要依赖于相对定量。通过LC-MS的绝对定量为 这是困难的，需要同位素标记的标准品和仔细的校准。同位素标准品价格昂贵， 很难获得。因此，可以通过靶向LC-MS-MS常规定量的代谢物的数量 方法一般小于500。另一方面，相对量化更容易， 可以使用“病例”和“对照”之间的峰值强度比较来相对量化数千个 不费吹灰之力就合成了。然而，相对量化有许多局限性和许多问题。在 特别是，相对值不能在实验室之间、平台之间、甚至在稍微分开的平台之间进行比较。 同一实验室内的时间段（批次效应）。这使得相对量化从根本上“不公平”， 数据共享或再现性的观点。此外，相对量化仅适用于某些有限的 实验设计（病例与对照）和相对值绝不能用于临床、法律的或工业试验 设置.这限制了非靶向代谢组学的应用，“仅供研究使用”。如果非靶向 代谢组学要想超越实验室，进入主流，它需要发展强大的， 无标记的定量方法，可以在不同的样品，跨平台，跨实验室和 穿越时空挑战是如何在没有同位素标准品的情况下通过LC-MS进行代谢物定量？ 幸运的是，最近出现了一些新的发展和新的想法，将实验性的 和计算方法，这表明有可能进行准确的代谢物定量， 无同位素标记标准品的非靶向LC-MS代谢组学。我们的目标是实施、测试和完善 这些方法，特别是极性代谢物，并使他们提供给所有感兴趣的CIDC成员。

SUMMARY The primary focus of the NIH Compound Identification Development Cores (CIDC) is to use untargeted metabolomics to not only identify novel metabolites but to facilitate and improve the identification of known metabolites. Furthermore, the CIDC is mandated to promote the accuracy, reproducibility, and interlaboratory comparison of metabolomics data. One way of promoting reproducibility, improving comparability and enhancing the confidence of metabolite identification is to improve metabolite quantification -- especially for untargeted metabolomics. Indeed, as frequently shown by untargeted NMR studies, knowledge of the concentration limits of a particular metabolite can “rule-in” or “rule-out” a tentative identification. For instance, if a metabolite signal is tentatively identified as kynurenic acid, but the measured concentration is determined to be 100X times more than normal, then that tentative identification must be incorrect and thus, “ruled out”. Traditionally compound quantification in metabolomics (especially absolute quantification) has been limited to targeted metabolomics while untargeted methods have largely relied on relative quantification. Absolute quantification by LC-MS is difficult and requires isotopically labeled standards and careful calibration. Isotopic standards are expensive and difficult to obtain. As a result, the number of metabolites that can be routinely quantified by targeted LC-MS- based methods is generally less than 500. On the other hand, relative quantification is much easier and it is possible to use peak intensity comparisons between “cases” and “controls” to relatively quantify thousands of compounds with little effort. However, relative quantification has many limitations and numerous problems. In particular, relative values cannot be compared across labs, across platforms, or even over modestly separate time periods within the same lab (batch effects). This makes relative quantification fundamentally “unFAIR” from a data sharing or reproducibility perspective. Furthermore, relative quantification only works for certain limited experimental designs (cases vs. controls) and relative values can never be used in clinical, legal or industrial test settings. This limits the application of untargeted metabolomics to “research-use only”. If untargeted metabolomics is ever going to expand beyond the lab and into the mainstream, it will need to develop robust, label-free quantification methods that can work across different samples, across platforms, across labs and across time. The challenge is how to perform metabolite quantification via LC-MS without isotopic standards? Fortunately, there have been a number of recent developments and novel ideas that integrate both experimental and computation approaches that suggest it may be possible to perform accurate metabolite quantification via untargeted LC-MS metabolomics without isotopically labeled standards. Our goal is to implement, test and refine these methods, specifically for polar metabolites, and make them available to all interested CIDC members.